petitcomputerfandomcom-20200214-history
Petit Computer Wiki:Chat/Logs/26 October 2013
05:43 Hye 05:43 Do you want to see some pattern recognition? 05:43 GimmeMoreCoinz? 05:49 hi! 05:49 I didn't think you'd be there. 05:49 I was looking at your video. 05:49 Oh cool 05:49 So you already saw it lol 05:49 mhm 05:49 I'm impressed. 05:50 What level of math (highest) did you apply to do it? 05:50 Thank you! I'm rather impressed myself, which is rather vain lol 05:50 Errr well, I'm in a graduate level Pattern Recognition class, so there's some of that in there 05:50 And let's see, some basic euclidean distance (the distance formula) 05:50 Errr vector angle calculation. I think that's pre-calc? 05:51 *angle difference 05:51 Data downsampling... I don't know where I got that from lol 05:51 And scaling, which is basically algebra 05:52 That's it 05:52 I got a pre calculus book for dummies at the library today 05:52 xD 05:52 because I began thinking.. damn.. I'm gonna need that for my AI eventually. 05:53 -.- also printed out the 5000 most common words in the english language. 05:53 Probably. Pre-calc turned out to be more important than calculus in mathematics applications 05:53 Because I took algebra and geometry I think I could get pre calc 05:53 what about the pattern recognition stuff. 05:53 are there tutorials on that stuff available online to? 05:54 Errr well, you don't really need that. I actually just used minimum distance, which is basically just calculating how "different" the given number is from the training data 05:54 I mean, it helped to understand the concepts, but you're really just comparing weighted values 05:54 so you used a neural network? 05:55 ABSOLUTELY NOT! Petit Computer would die lol 05:55 XD 05:55 well there goes my plans 05:55 I used a simple "minimum distance" classifier. Not really exciting 05:55 I wanted to use a neural network for npc character AI in a game 05:55 Errr I mean, you could try 05:55 it's hard to represent neural networks in BASIC 05:56 But it's thousands of calculations per action, which can lead to some crazy wait 05:56 aha 05:56 Are you up late when you shouldn't be? XD I't's my weekend and I already had an 8 hour nap. 05:56 i'm in a course right now. 05:56 It's only 2 here 05:57 only 2. ^.^ 05:57 2am on a Friday 05:57 It's 11 pm here. 05:57 *Saturday I guess 05:57 I'm in a sort of.. rut. 05:57 You don't need to use neural networks for AI 05:57 I know. 05:57 And I'm sorry about that 05:57 I'm trying to figure out how to progress my AI 05:58 The idea is it's language agnostic.. so programming a match case for each word. or having set grammar rules would be difficult to keep the scope of it's intended functionality.. but right now it also has a high grammar error ratio.. very high o.0 05:58 A neural network would be the best for AI that is able to adapt, but that's something for desktops to handle, not tiny PTC 05:58 It's for desktop :3 05:59 The way it works right now is it remembers input, and then refers to the output used for that input so it can copy the user's responses to it's own responses if the user uses it's response the user gave. 05:59 Me: Hello 05:59 Ai: Hello. 05:59 Me: How are you? 05:59 05:59 Session 2: 05:59 Me: Hellow 05:59 AI: How are you? 05:59 I see 05:59 brb. 06:01 I don't have alot of experience working with AI. So it's difficult for me to come up with methods for dealing with that sort of data and Ive never constructed a working neural network before either. 06:01 Well, neither have I lol 06:01 What would you suggest I make my first step? It's meant to be a language agnostic continuous learning AI that is entirely adaptive. 06:01 XD 06:02 Errr I have no idea. Why not try to make a neural network to classify small things first? 06:02 but yeah you also said I wouldn't need a neural network which is an interesting point because the current version isn't using one. 06:02 that's a good idea. 06:02 Like... make a neural network for shape recognition or something 06:03 that'd be cool. 06:03 brb again I need to move a lamp. and get headphones. 06:03 There's a lot more to neural networks than training though. There's a lot of data transformations in order to get the image to fit the training samples 06:03 mm.. 06:03 data transformation 06:03 I see so you use a scalar for the inputted numbers 06:03 then you do the comparisons 06:04 What? 06:04 I use a vector 06:05 Where do you start your calculation? Is it at the start of the first touch position? 06:05 I performed some data manipulation myself on the number recognizer, and I'm not even using a neural network 06:05 Where do I start? 06:05 yeah that's really interesting. 06:05 Hmm, well, here's how it works: 06:05 See I wanna learn to do that same kind of stuff. 06:06 First, I read in all the points in that were drawn. Like every single one. 06:06 whoa. 06:06 Then I downsample the points so that each number is only stored as 20 points 06:07 I would have thought there might be a shortcut. o.o 06:07 That's the shortcut lol 06:07 I would have thought it would be something like only storing every fifth or tenth point 06:07 and then scaling. 06:07 I suppose doing that your error rate would go up 06:08 Yes, so downsampling is better. Notice that the numbers on the left still look like the numbers on the right? That's because I did some super nice downsampling lol 06:08 Oh wow. 06:08 I shifted the points around so that it still follows the conour of the number basically, instead of just leaving out points 06:08 *contour 06:08 I'm developing a drawing system/AI.. 06:08 Anyway, after I have the points, I scale the image to a fixed height 06:08 oh 06:09 A drawing system? 06:09 Yeah I'll tell you when you've done explaining I didn't realize there was more. 06:09 Lol there's a lot mor 06:09 After I scale the image, I shift it so that they all start at 0,0 06:09 Basically, this is called "normalization" 06:09 wait... 06:10 Yes? 06:10 do you mean, you aren't using that coordinate system to start with? 06:10 Which one? 06:10 0, 0, screenwidth, screenheight as coordinate range where 0, 0 is the top left orner. 06:10 corner* 06:10 Oh, I mean I could have shifted first 06:11 I mean the starting point for the drawing will be 0,0 06:11 I don't quite understand how the point's x and y changes. 06:11 When you read it in, it'll be tchx and tchy 06:11 right. 06:11 But then, if someone draws on the corner instead, it won't be recognized correctly 06:11 wierd. 06:11 So I shift them all so they all start at 0, no matter what 06:12 so that fixes the recognition problem? 06:12 Err it fixes the problem of drawing numbers in different spots 06:12 Right. 06:13 Anyway, now that all the numbers are the same size and in the same place (and stored as vectors of 20 points), it waits for another number 06:13 It processes the number in the exact same way, so it's the same height and in the same place as the "reference" numbers 06:13 Now comes the actual recognition part! 06:14 I first calculate the distance between each point between the given and each number 06:14 This is basically "how much does the number LOOK" like another? 06:14 You know how to calculate distance right? 06:14 Sort of. 06:14 It would be something like getting the points AB and CD 06:15 SQRT((X2-X1)^2 + (Y2-Y1)^2) 06:15 Then doing some math on each set of points to get their relative distances. 06:15 o.0 06:15 right I remember learning that in algebra/geometry but I had forgotten the formula 06:15 it was more like. 06:16 Lol that's ok 06:16 Anyway, now that we have the distance between our points and the other numbers, it's time to weight each point 06:16 For instance, a 3 may look suspiciously like a 1 if drawn without too much curve 06:16 A^2+B^2=C^2 06:16 Right, but you want C^2 06:17 Sorry, you want just C 06:17 XD 06:17 So SQRT 06:17 oh ok. 06:17 And A and B are lengths, but our X and Y are points, so we need the length between our point (X1) and their point (X2) 06:18 Anyway, let's say we have our distances 06:18 Now we need to weight the distance based on the direction of our stroke 06:18 The distance by itself will actually recognize numbers just fine, but I want something stronger 06:18 Ah. 06:18 I'd want something stronger myself to.. 06:19 I'll tell you more about that idea later. 06:19 To weight the distance, I simply calculated the angle between the current stroke and the stroke of the reference numbers 06:19 For instance, if you're drawing a six, it looks suspiciously like a 0 when you start 06:19 oh right.. and because your representing curves as linear lines AB and CD it would just be like.. still a straight line. 06:20 with an angle? 06:20 However, as soon as you start to veer off course to make the little loop, the weight starts kicking in and instead of getting a small distance between your number (6) and 0, now it's very large because you changed stroke 06:20 hmm. 06:20 And we just pick the number with the smallest weighted distance. 06:21 that's pretty nice. 06:21 I could draw it better lol 06:21 Yeah. 06:21 I'll most likely read some of your code to try to understand the concept more. 06:21 when you release it. 06:23 The drawing Ai I want to create would basically have a small database, and a draw board. 06:23 I could release it now, it's just got a bit of a sudoku wrapper around it lol 06:23 It would calculate not just angles and lines but types of curves and what types of curves are connected. 06:23 It would sort smaller sets of strokes ( Pen off pen up pen on) 06:23 and it would take note of their relation to other strokes. 06:24 then it would basically try to create a math formula that summarizes the drawing. 06:24 Ah, that's very difficulty 06:24 so in other words (it would try to convert to a graphing equasion) 06:24 *difficult 06:24 of some sort. 06:24 so it would figure out how sharp the curve is 06:24 if it's a bezier... etc 06:24 how many similiar curves are nearby 06:24 and also use some visual image recognition 06:24 The cool part is 06:25 it could then find art in it's database that is closely similiar 06:25 it could also try to generate the average between two or more pieces of art 06:25 so it would average out the strokes in order/location tec 06:25 to create something unique.. 06:25 and with less randomness it could recognize to a point what it is drawing itself using feedback 06:25 That would be neat 06:25 so eventually if it used a neural network 06:25 with some more complex algorithms.. 06:26 it could learn to draw from people. 06:26 it would be really abstract though 06:26 but it would be intruiging 06:26 then the third idea for it would be to have it rate you based on other samples.. and tell you "you used this many types of curves this many straight lines this many " and so on 06:26 so you get statistics on how you are doing your art and it would actually be able to find other artists who do the same types of things. and group your style 06:27 This is such a ridiculously hard problem, that it's probably someone's thesis or research topic 06:27 I'm ambitious. 06:27 and that's the whole reason I grabbed pre calc for dummies today 06:27 xD 06:28 Eventually combining my two AI projects into one would be crazy. 06:28 an AI that can speak natural language and do those tasks, and relate them. 06:28 Lol you know that's not even done yet, right? 06:28 yes 06:28 Well, good luck 06:29 I'm working on things no one's done. 06:29 Oh, here's the graphical representation of what I was talking about 06:29 thanks I'm going to need to be studying alot of math and reading alot. it's not easy to even fathom it. 06:30 Oh randomous I have a question for you, if an AI doesn't use a neural network and it doesn't actually have a true comprehension of language but gives the illusion that it does, is it just as valid as one that has a true comprehension of langauge if they perform similarily with no (humanly noticable) differences 06:30 Argh, I can't find the picture lol 06:30 Errr well, if it doesn't use a neural network, it's much harder to learn on the fly 06:30 It's generally not possible 06:30 You'd be interested in this then... 06:31 Well, I don't know, I'm not as ambitious as you lol 06:31 https://db.tt/07x4WWl7 06:31 Let me know when you see it 06:32 The red is a stored zero, the two is our drawn two 06:32 http://imgur.com/vRoma8d 06:32 okay and let me know when you read that. 06:33 I did 06:33 But it's going to be hard to get it to say different things if it starts to say something inappropriate 06:34 Unless you use something like a neural network 06:34 You may even be implementing one without knowing it lol 06:34 You think so? 06:34 Want to see some of my code? 06:34 No, it's OK 06:34 I'll just explain what a neural network does in general 06:34 Or do you already know? 06:36 I kind of know but you could explain. 06:37 OK, so basically a neural network works by looking at a set of old situations to figure out how to translate input into output 06:37 A learning neural network will add input to the set of old situations as it comes across them 06:38 In order to choose an output, you simply pick the situation that best matches the input, and chose it's output 06:38 Sometimes there's multiple things to do, in that case, you weight each one depending on how many times you've come across it, or perhaps a combination of that and your own function 06:39 (which can be anything) 06:39 This is messy I had a point in time where I was trying to brute force a few things I really wanted it to know. :/ But I felt it was partly cheating for a "intended to be language agnostic learning AI" 06:39 http://pastebin.com/XdCqn3QF 06:39 You might be surprised x.x 06:39 That's how the neural network learns, it realizes that people have said "Hello" after "Hi" like a billion times, to it figures that's the best one to do 06:39 and you're probably right I'm probably in some way representing a neural network. 06:39 o.0 06:39 uhm 06:39 I'm pretty much doing exactly that... in a sense 06:40 I have a system that sort of does that 06:40 Then you're making a neural network 06:40 That's all it is 06:40 o.0 06:40 cool so I'm actually on the right track. 06:40 Really, all the fancy stuff floating around it is just extra 06:40 oh it is? 06:40 like weights, and such. 06:40 I had some sort of weighting system in it. 06:40 but it's based on 06:40 Well, it's usually a good idea to have something to differentiate input 06:40 Ok in order to try to simulate natural language it checks words to try to figure out the most commonly used ones, and those get used more often during random sentence generation which is just a cheap way to compensate for a non response. 06:41 it starts out not knowing any thing 06:41 the reason is cause what it knows has a corresponding array "input responses" 06:41 I made a simple java version that didn't have the learning. 06:41 javascript. 06:41 And I had to manually input the responses and inputs into arrays. 06:42 Why not look at cleverbot? 06:42 It's basically exactly what you're making 06:42 You can just tune it to how you want it made 06:42 http://www.cleverbot.com/ 06:46 actually some if it is based on the documentation of how to program cleverbot 06:46 XD 06:47 Lol 06:48 Yeah. I sat and read the entire documentation atleast twice when working on my AI and my friend told me You know cleverbot is open now to people to use 06:48 I was like WHATT?! 06:48 o.0 06:48 It's been use for years lol 06:48 I learned a /few/ neat tricks about referencing. 06:48 but the thing this AI can't do yet is use meta tags that represent abstract structure. 06:48 For instance I 'verb' you are 'verb' 06:48 response 06:49 "Oh how interesting that you think you 'verb' are 'verb2' 06:49 but I don't really like that approach which is why I wont touch AIML 06:49 i just.. have a disdain for it 06:49 I don't really know a lot about this stuff 06:49 oh. 06:49 Language stuff, I mean 06:49 Yeah. 06:49 Well 06:49 With AIMl you can program it to use a keyword. 06:49 I can do the math and stuff for number recognition, but I just don't have the mind for language 06:49 XD 06:49 oh 06:50 I'm the opposite. 06:50 I'm poorer in math and my literacy is high. 06:50 that's part of why I can program because I memorize syntax very well. 06:50 We can't be good at everything 06:50 no definitely not 06:50 I got a 9/9 on my CAAT Level C test 06:50 in grammar sensitivity. 06:50 I don't know what that is 06:50 I suppose not lol 06:50 it's just basically a standard knowledge and aptitude test 06:50 I figured lol 06:50 that shows where your strengths are in academic subjects 06:51 I'm above average (highest scoring) 06:51 in language. 06:51 which even surprised me x.x 06:51 it kinda makes sense to me now why I confuse people when I talk.. 06:51 Lol well, you should definitely pursue something in that area then 06:51 yeah and I was tired on testing day. 06:51 Canda! 06:51 so that's actually what would be a lower score for me! 06:52 Pfft Canada 06:52 I can't even spell. That's what it is 06:52 Funny I didn't catch your error 06:52 the first time 06:52 :/ 06:52 So you're in Canada... I guess the CAAT is similar to the SAT here 06:52 probably? I'm not sure. 06:53 usually it's timed and colleges give it to you. 06:53 it's to do with aptitude I think. 06:53 some of the scoring is very sensitive and it's based on percentile. 06:53 "Standardized Aptitude Test". It's what you take so that colleges know how good you are 06:53 Ah. 06:53 Then probably. 06:54 Would you say it's normal to sometimes wish I was better at math and such. Being a programmer. 06:54 Whhops, it's "Scholastic", not standard lol. Goodness, see how bad I am at language? 06:54 Oh well :3 I got what you meant atleast...although there is a huge gap in the meanings there. 06:54 The outcomes the same. 06:54 Sure. Programming has a strong math background, but you don't necessarily need to be a math master 06:54 My head is full of cobwebs right now. 06:55 I didn't sleep well all week. 06:55 Well, get some sleep lol 06:55 I'm awake. XD Because I had a nap. 06:55 Totally random, but do you speak French? I don't, but I was just wondering because Pokeyoshi does 06:55 I don't speak french no. 06:55 I know like four words or phrases. 06:56 Hmmm I think Pokeyoshi said he was from Quebec, which explains the whole French thing 06:56 I probably am more fluent in japanese than french which is wierd... 06:56 Nah, not weird 06:56 Yeah not everyone in Canada learns french. 06:56 Or they do but they don't really try to remember. 06:56 It's not like you have to know French to be in Canada lol. And I know everyone doesn't speak French in Canada lol 06:56 Where I'm living there's alot of natives. as well. 06:57 I'm actually very interested in language.. 06:57 I noticed 06:57 but I have a hard time learning.. sometimes. 06:57 My head is always crammed full of other crap 06:57 XD 06:57 Hmmm take some time off and relax 06:57 I'm also interested in finding new ways to mathematically represent problems. 06:58 You told me. 06:58 oh. I guess I did. haha. 06:58 THAT requires a lot of math lol 06:58 Yeah. 06:58 Math is going to become a huge field for me to continue learning. 06:58 but sorta...pushed to the side of everythin gelse. 06:58 And then there's my music interests.. >.< I want to get good at keeping rhythm and percussion. But simple things first.. 06:59 then composing. I've always been ok at making melodies. 06:59 >.< 06:59 I'm somewhat OK at musi 06:59 I played Clarinet for... let's see 06:59 10 years 06:59 Yeah that's incredible. 07:00 And played the recorder for a year recently (but just for fun) 07:00 xD 07:00 I've messed with pianos alot. 07:00 I worked at Medieval Times, and I was a villager, so I figured I'd play the part and play some music 07:00 I was visiting my friend and I walked into a music store with him and began playing the mario world theme. 07:00 Lol that's awesome. I've always been terrible at piano 07:00 I'm better at piano than anything else. 07:00 Woodwind instruments are my thing 07:00 I have coordination issues. 07:01 but I can type almost 180 words per minute averaging 160 on a good day. 07:01 so that explains where that comes from 07:01 Errr I'm pretty sure I type slowly 07:01 if I can push buttons to play music it works for me. 07:01 strings? ughh.. I tried alot. I can sorta play some basic guitar but nothing like chords. 07:02 and I absolutely horrid at rhythm instruments or percussion 07:02 Which sucks cause I wanna get into tracking. 07:02 I'm not bad at percussion, but that's probably because of the whole 10+ years in music thing lol 07:03 mhm 07:03 I was in my high-school's drum core, so that was cool 07:03 I was walkign home with my 3ds messing with the drumming feature. all day today. for like two hours. 07:03 trying to train my brain to keep a basic time. 07:03 part of the issue is getting my fingers to move when my mind wants them ot. 07:03 to* things like that. 07:03 i notice sometimes if I try to go too fast my finger just freezes up or twitches. 07:03 and I mess up 07:04 I was actually thinking of trying to make a rhythm game.. even though I tend to hate them (but guitar hero is different) 07:04 I played Guitar Hero 07:05 I was actually ranked #8 on the Wii version of Guitar Hero World Tour for a while 07:05 ...but that's neither here nor there lol 07:06 Anyway, I have to go. I'm really sorry for cutting the conversatino short, but I have to go to bed 07:06 Thanks for chatting for so long 07:09 It's okay. 07:09 Thanks for talking with me to. I almost always enjoy conversations with you. 07:10 Lol almost 07:10 I'm sorry if I ever put you off 07:10 You don't really it's just sometimes I'm not in the right state to talk :3 07:11 I understand 07:11 night' 07:11 Well, I hope you have a good night! 07:11 You too 07:11 Sleep good. 07:11 Thanks, you should sleep at some point too. 07:12 mhm 08:14 testing 08:20 person do you know a list of what graphic colors are there numbers? i don't want to go through 255 colors for one color 08:20 i know 1 is grey, 67 is red, and 15 is white 08:20 thats it 08:32 wait person i can use grped to make the textures? 08:32 yea 08:32 :0 08:32 >.< 08:54 ... 08:54 Hello. 08:55 ... 08:55 There is a section in the PTC manual that shows all the colors. 08:55 HERRO AGAIN!! 08:56 ... 08:56 Hi. 08:57 Does vsync do the exact same thing as wait? 08:58 I'm finally back on the wiki. I think I've been gone over a week and a half. (That's longer than Gimmemorecoinz was gone, and that felt long to me.) 08:59 no 08:59 vsync stops graphics for a set amount of frames 09:00 wait stops everything 09:00 Oh. 09:01 No vsync makes the BUTTON command register at the same time as.screen refrwzh, otherwise you would have player sprite zooming offscreen with almost no hold time, but wait pauses everything, and I think vsync only stop some things. 09:02 Woah. Sorry for the typing. Using different device. 09:02 Just try replacing vsync with wait. 09:05 It doesn't seem to change anything. 09:06 no vsync does graphics 09:06 do this: 09:06 @LOOP 09:06 CLS 09:06 PRINT "HELLO" 09:06 VSYNC 1 09:06 GOTO @LOOP 09:06 it doesnt flash 09:06 so it has to affect text/graphics 09:06 Use wait now. 09:07 Same results. 09:07 Ehh. I only ever used for spritez. 09:07 because the interpretation time is so little 09:07 it doesnt make a difference 09:08 Trying to find asterisk... 09:08 Whaat? 09:08 but if i put WAIT 2 and VSYNC 2 in minecraft ds's code 09:08 itd make a huge difference 09:08 Hmm. 09:08 What if you used gline for mcds? 09:09 Instead of gfill? 09:09 cause then i'd have to use GPAINT to fill it in 09:09 +imma use this cuz stupid tablet keyboard has no asterisk. 09:09 and 4 GLINEs to make a square 09:09 its be so painfullt slow 09:09 Yeah but it would better. 09:09 even if i used GBOX+GPAINT itd be to much slower 09:09 Yeah but it would look better. 09:10 Sorry. 09:10 and it wouldnt look bette 09:10 better* 09:10 itd look the same 09:10 just slower 09:10 I think it would. 09:10 I made an example. 09:11 Just a minute. 09:14 itd physically we the same 09:14 be*\ 09:15 person 09:15 what 09:15 Back. 09:15 did u see my pm 09:16 yes i did 09:17 also i made a night sky if u want it 09:18 cool 09:18 making halloween texture packs r fun 09:18 is* 09:18 may i have the night sky? 09:19 when i release the halloween texture pack :3 09:20 no 09:20 i need now 09:20 XD 09:20 http://img585.imageshack.us/img585/3999/81nk.png 09:20 ok 09:20 Gline 3d. 09:22 8x8 texture. 09:22 what the heck 09:22 is this 09:23 yes but thats not minecraft ds 09:23 It an example of using gline instead of gfill. 09:23 I can't do raycast. 09:24 if you use GLINE with raycast, itd be slower and look the same 09:24 Did you look at it? 09:24 i agree that this looks good, but i kinda cant just change the entire engine of minecraft ds 09:24 and i did look at it 09:24 It draws it 50 times. 09:26 Maybe theres a way to do it without raycast. 09:27 im not changing the engine of minecraft ds 09:27 it works how it is 09:27 Ok. Thats cool. 09:27 Trying to improve performance? 09:28 Its pretty amazing what you did. 09:28 you can improve performance already 09:28 in the mcds chat 09:28 type: 09:28 /quality 09:28 enter 09:28 12 09:28 enter 09:28 have fun with lagless minecraft ds 09:28 I know that, I was asking what you were talking about. 09:29 Scroll up. 09:29 Haha, I did 99 once. 09:29 lol 09:29 thow fast 09:29 how* 09:29 It's... IDK, stupid? 09:30 lol 09:30 Pretty much no draw lag, because it didn't have to do much 09:31 :0 09:31 quality at 12 is lagless 09:31 I'VE BEEN LIVING A LIE 09:31 person 09:32 did u get the grp 09:32 But it's so stretched out that one block takes up the whole screen as about six pixels from 8 blocks away. 09:32 yeah... 09:32 On 99 09:33 person! 09:34 Couldn't you theoretically make the textures 4x4, and it would be permanent q12? 09:34 nope 09:34 4x4 textures are for chumps 09:35 q12 is a 12-pixel gap between columns 09:35 not the same as 4x4 textures 09:35 minecraft looks HD at quality 1 but i can barely move at how slow it is... 09:36 yea 09:36 I meant more of similar in speed and ability to tell what a block is. 09:36 textures dont change the speed much at all 09:36 i did 1x1 textures once 09:36 person did u get the night sky grp 09:36 and it wasnt much better than 16x16\ 09:37 and yes i did, thank u levi 09:37 do you like it 09:37 ? Then why were the textyres changed to 8x8 version? To test? 09:37 ikr 09:38 Oh, slow typing on this screen. 09:38 yes to test 09:40 ... 09:40 well... 09:41 ima go make the texture pack persson bye 09:41 bye 09:44 Did anything interesting happen on the wiki in the past week and a half? 09:45 gimme kinda came back 09:45 i released mcds b1.7 and 1.7.1 09:45 I totally made the best program ever 09:45 (not really) 09:45 lol 09:46 Sudoku? 09:46 Just the number recognition part. You saw the video I guess? 09:47 I'm not sure if I like sudoku or not. And yes, that was pretty cool. 09:47 Thanks 09:47 Once I get the code all cleaned up, I'll release it properly so that you can make stuff with it 09:48 ...not that it's all that useful though 09:48 How many closeness hits does it need to guess a number? 09:49 What do you mean? 09:50 It downsamples each number to 20 points, scales it so that they're all the same size, then shifts the numbers so they're all in the same space 09:50 How many (pixels?) Need to be similar to the original for it to... I can't think of the word. Match? 09:51 Nvm on that 09:51 Finally, when you draw a number to recognize, it checks it's 20 points against the 20 points from all the rest of the numbers by computing the distance 09:51 Lol sorry 09:52 Mmhmm, I was think of something like that with the first part reversed. Which would less effecient. 09:52 The distances are actually weighted by the likeness of stroke to make it even more accurate 09:53 For instance, if two points are very similar, but the stroke is in the opposite direction, it will be very heavily weighted (a lower number means a better match) 09:54 Ugh, my brain is really sluggish today. It took about 30 seconds for it to remember "efficient". 09:54 Lol it's Saturday, it's understandable 09:57 Also, because of... gdhvvdj , while I was gone, I had to use the alternate wiki activity section/tab/thing to see stuff and I saw all the green and red things that look like they are tracking positive xontributionism. What exatly are those? 09:57 What? 09:57 Lol xontributionism. Oops. 09:57 Green and red things? 09:57 Numbers, sorry 09:58 Ohh 09:58 That's how many characters on the page were changed 09:58 For instance, a red 200 means that 200 characters were removed 09:58 A green 1000 means that 1000 characters were added 09:59 Ohh... well, a lot of the red ones I saw were next to posts with bad grammar amd spelling XD 10:00 That's probably because the bad stuff was removed lol 10:02 And a lot of your comments and stuff had high positive numbers. 10:03 That's because I'm always appending stuff to them lol 10:03 I'm not good at settling my mind on a comment 10:05 We had our first real geometry test last wednesday, and the highest score was a B minus, and most of the smart people got D's or C's,, so we had a retake yesterday. 10:06 Lol sounds like a hard test 10:08 The really bad thing was that it was mostly careless mistakes lime not distributing the negative or writing sameside instead of alternate interior. 10:08 Hey 10:08 Lime? I wish I had a real keyboard. 10:08 Lol ime 10:09 . 10:09 *lime 10:09 I have to go, so I'll see you all later 10:09 Fail 10:09 Hey, Randomous 10:09 oh bye,,, 10:09 Hey 10:09 Ok bye. 10:09 *Cries 10:09 Sorry, I just have a lot of work to do 10:09 ...Well bye 10:11 I'm not sure if I want to sleep or eat something. 11:35 we all wish this was true 11:36 http://prntscr.com/1zy29q 11:36 The prices XD 11:36 Imma keep messing up webpages. 11:37 & taking pics of them for distribution becuase it is funny XD 11:48 COINZ! 11:48 OH SPARX! 11:48 hi 11:48 where have u been?'w' 11:48 mee? 11:48 Yeah 11:48 I slept 11:48 xD 11:48 XD 11:48 I don't feel very awake still despite being up for over five hours already 11:49 sparky what are you doing 11:49 ? 11:49 oh 11:49 messing with webpages 11:49 http://prntscr.com/1zy29q 11:50 m 11:50 and viewing the wikia 11:50 read the prices 11:50 Im messing with them to ake myself and whoevers here get a lol 11:50 *make 11:50 I need help coming up with game ideas... 11:51 you'll be surprised to hear that petit smash bros basically needs another re hash entirely. 11:51 -wut 11:51 petit smash bros has to be re coded from the ground up entirely a third time 11:51 What happened!? 11:51 nothing. 11:51 D: 11:51 it just doesn't meet my requirements anymore. 11:51 Then why's that?:/ 11:52 What are your requirements? 11:52 quality. 11:52 smoothness 11:52 a good core game engine 11:52 Omg,stop being so hard on yourself,The game is going great 11:52 Yoshi:COINZ!>:D 11:52 I'm thinking of changing petit smash bros to use vector collision and graphical slices for sprite parts. 11:52 * 11:52 no 11:52 i'ts not going great it's terrible and needs a re work 11:52 xD 11:52 Yoshi: 11:52 Coinz! 11:53 I think the game is great 'w' 11:53 I'll probably just do petit smash bros over on pc first. 11:53 Coinz! 11:53 hi yoshi 11:53 see!X3 11:53 xDDD 11:53 Hai 11:53 Hi 11:53 I have some ideas.. 11:54 if I can do per pixel blending... and stuff 11:54 and add alpha blending of sprites to an engine 11:54 I can do some pretty good graphical effects 11:54 m...hmhm 11:54 that's perfect... 11:54 I thought of a new menu design. 11:55 The character select will be circles. 11:55 like a rotating selector 11:55 "." 11:55 hi twin 11:55 why was i not invited to this party 11:55 twin!:D 11:55 is this rush hour on ptc wikia or what 11:55 yeah 11:55 I guess all noticed ur in chat 11:55 guys 2013 10 26